How would you structure an experiment to determine whether a new messaging cadence improves conversion?
